Conference Keynote by Jen-Hsun Huang, founder and CEO of NVIDIA
Tuesday, November 15, 8:30AM

Kicking off this year's conference is the keynote by Jen-Hsun Huang, founder and CEO of NVIDIA. An on-demand version of the keynote will be available soon, please check back here.

GPU Technology Theater @ SC11
Monday, November 15 – Thursday, November 17 during exhibition hours | NVIDIA Booth #2719

Click on the Theater tab for more details, or Download the PDF schedule.

To view or download presentations from the GPU Technology Theater at SC11 please visit GTC On Demand.

CUDA Research Fast Forward
Monday, November 14^th, 20:00PM|NVIDIA Booth

Presented by the Pasi Fellows, Moderated by Lorena Barba, Boston University

Albert Sidelnik, University of Illinois Urbana-Champaign
Christopher Cooper, Boston University
Trevor Gokey, San Francisco State University
Olesiy Karpenko, University of Illinois Chicago
Anush Krishnan, Boston University
Simon Layton, Boston University
Ying-Wai Li, University of Georgia
Britton Olson, Stanford University,
Juan Perilla, University of Illinois Urbana-Champaign
Benjamin Payne, Missouri University of Science and Technology

For a detailed listing of the CUDA Research being presented, please Download the PDF here.

Tutorial: High Performance Computing with CUDA
Monday, November 14, 8:30AM-5:00PM

*See below to download PDF copies of the presentations.

This tutorial will introduce CUDA to the supercomputing audience and motivate its use with traditional HPC examples. We will first teach the basics of CUDA programming with step-by-step walkthroughs of code samples, then review the main optimizations techniques, and describe profiling and tuning best practices to maximize performance. While CUDA C and CUDA Fortran will be used for illustration, the concepts covered will apply equally to programs written with the OpenCL and DirectCompute APIs. Finally, we will close with case studies from academia and industry.

Download Introduction PDF
Download CUDA C Basics PDF
Download CUDA Fortran and Libraries PDF
Download Performance Optimization PDF
Download Multi-GPU Programming PDF
Download Unrolling Parallel Loops PDF
Download CUDA Accelerated Monte Carlo for HPC PDF

GPU Technology Theater @ SC11
Monday, November 14 – Thursday, November 17 during exhibition hours | NVIDIA Booth #2719

The GPU Technology Theater hosts talks on a wide range of topics on high performance computing. Open to all attendees, the theater is located in the NVIDIA booth and will feature industry luminaries, scientists, and developers.

Download the PDF schedule

To view or download presentations from the GPU Technology Theater at SC11 please visit GTC On Demand.

The GPU Technology Theater talks are now a part of GTC On-Demand. Presentations will be made available after SC11 at http://www.gputechconf.com/page/gtc-on-demand.html.

GTC On-Demand gives you archival access to the world-class education delivered at GTC and is an essential resource for the scientists, engineers, researchers, and developers who rely on GPUs to tackle enormous computational challenges. Visit GTC On-Demand today and explore and learn from the best and brightest minds working in High Performance Computing.

Developer Demos in NVIDIA Booth

NVIDIA is hosting developer demos that will help you accelerate your code on GPUs. Demos feature CUDA C, libraries and directive-based solutions. Can't make it to the NVIDIA booth? Check out the demo videos online at NVIDIA's Developer Zone.

Tutorial: High Performance Computing with CUDA
Monday, November 14, 8:30AM-5:00PM

This tutorial will introduce CUDA to the supercomputing audience and motivate its use with traditional HPC examples. We will first teach the basics of CUDA programming with step-by-step walkthroughs of code samples, then review the main optimizations techniques, and describe profiling and tuning best practices to maximize performance. While CUDA C and CUDA Fortran will be used for illustration, the concepts covered will apply equally to programs written with the OpenCL and DirectCompute APIs. Finally, we will close with case studies from academia and industry.

Panel: HPC Analyst Crossfire
Friday, November 18, 1:30PM-3:00PM | TCC 101

Addison Snell, a leading HPC analyst with Intersect360 Research, will moderate this lively panel discussion, in which he asks visionary leaders from the supercomputing community to comment on forward-looking trends that will shape the industry in 2012 and beyond. An audience Q&A for the panelists will follow the live recording session.

Paper: Fast Implementation of DGEMM on Fermi GPU
Wednesday, November 16, 10:30AM-11:00AM | TCC 303

The GPU is offering more than an order of magnitude speedup of peak floating-point computing over conventional processors. In this paper we present a thorough experience on tuning double-precision matrix-matrix multiplication (DGEMM) on the Fermi GPU architecture. We choose an optimal algorithm with blocking in both shared memory and registers to satisfy the constraints of the Fermi memory hierarchy. Our optimization strategy is further guided by a performance modeling based on micro-architecture benchmarks. Our optimizations include software pipelining, use of vector memory operations, and instruction scheduling. Our best CUDA algorithm achieves comparable performance with the latest vendor supplied library: CUBLAS 3.2. We further improve upon this with an implementation in the native machine language, leading to a 20% increase in performance over CUBLAS. That is, the achieved peak performance (efficiency) is improved from 302Gflop/s (58%) to 362Gflop/s (70%).

BOF: High Level Programming Models for GPU Computing
Wednesday, November 16, 5:30PM-7:00PM | TCC 202

There is consensus in the community that higher-level GPU programming models based on directives or language extensions have significant value for enabling GPU programming by domain experts. Several efforts are under way to develop such models layered on top of standard C, C++ and Fortran through either standards committees or the introduction of proposed de facto standard solutions by large industry players. This BoF will explore and debate the merits of several current options and approaches for high-level heterogeneous manycore programming.

Exhibitor's Forum: Industry Adoption of GPU Computing
Thursday, November 17, 1:30PM-2:00PM | WSCC 611/612

Since its introduction just a few years ago, the high performance computing (HPC) industry has widely adopted GPU acceleration to solve extremely challenging problems. Industry's computational need is constantly increasing as large and complex computational problems become commonplace. We will discuss the driving forces behind the rapid adoption of GPU computing and explore its impact across various industries.

Posters on display during exhibition hours | NVIDIA Booth #2719

Adaptive Beam-forming for Radio Astronomy on GPUs
Vamsi Krishna Veligatla - University Of Groningen

BarraCUDA - a Fast Sequence Mapping Software using GPUs
Brian Lam - Cambridge University

Black Hole Astrophysics
Gaurav Khanna – UMass Dartmouth

Computer-Aided Nano-Design
Yan Wang - Woodruff School of Mechanical Engineering, Georgia Institute of Technology

Debugging Floating Point Implementations on GPUs
Devon Yablonski - Northeastern University

Efficient Dense Stereo Matching using CUDA
Ke Zhu - Technische Universität München

ExaFMM: An open source library for Fast Multipole Methods
Lorena Barba - Boston University

FEN ZI: GPU-Enabled MD Simulations Based on CHARMM Force Field and PME
Michela Taufer – Department of Computer and Information Sciences, University of Delaware

Flexible X-Ray Image Processing on GPUs
Matthias Vogelgesang - Karlsruhe Institute of Technology

GooFit for HEP Analysis
Karen Tomko – Ohio Supercomputer Center

GPU Accelerated Ultrahigh-Speed Real-Time Fourier-Domain Optical Coherence Tomography
Kang Zhang - GE Global Research

GPU-Based Molecular Dynamic Simulations Optimized with CUDPP and CURAND Libraries
Tyson Lipscomb - Wake Forest University

GPU Speedy Particles and Grids
Douglas Enright - Enright Labs

KILO Transactional Memory for GPU
Wilson Wai Lun Fung - University of British Columbia

LAMMPS with GPUs
Axel Kohlmeyer - College of Science & Technology, Temple University

Lattice QCD with GPUs
Andrei Alexandru – The George Washington University

Multi-GPU Computing
Jeffery Stuart - UC Davis

OpenFOAM Compatible Transient Solver for Incompressible Fluids on GPU
Lukasz Miroslaw - Vratis Ltd.

Parallel VLSI CAD Algorithms for Energy Efficient Heterogeneous Computing Platforms
Zhuo Feng - Michigan Technological University

Physis: An Implicitly Parallel Framework for Stencil Computations
Naoya Maruyama - Tokyo Institute of Technology

Plane Wave Pseudopotential Density Functional Theory Calculations on GPU Clusters
WeiLe Jia - Supercomputing Center of CNIC, Chinese Academy of Sciences

Porting and Testing DFTB Dynamics to GPU
Jacek Jakowski – National Institute for Computational Sciences

Protein-DNA Docking Via a Structure-Based Approach
Bo Hong - GeorgiaTech

Quantum Monte Carlo for Accurate Predictions of Correlated Electronic Systems
Jeongnim Kim – University of Illinois